Techniques for regression testing: selecting test case sets taylored to possibly modified functionalities

Regression testing is one of the most important and expensive activities of software maintenance. It involves testing the modified program to reveal faults introduced during maintenance and keeping the software´s level of reliability, as cheaply as possible. Two techniques of regression testing are discussed. The first is a selective technique (H.K.N. Leung and L. White, 1989) which identifies the modified structural attributes of a program (required elements) and selects a retestable test case set to exercise them. This technique is based on the potential-uses criteria family (J.C. Maldonado et al., 1992) and is implemented by a regression testing tool (I. Granja, 1997) working in conjunction with a testing tool (M.L. Chaim, 1991). The second technique comprises procedures for testing and regression testing activities. These procedures make use of the first technique to select a retestable set containing test cases capable of checking up on the potentially modified functionalities. The retestable set provides a good coverage of the structural required elements. Finally, empirical results show the effectiveness of the techniques